U.S. patent application number 12/065240 was filed with the patent office on 2009-09-03 for method and apparatus for correcting the output signal of a radiation sensor and for measuring radiation.
This patent application is currently assigned to PERKINELMER OPTOELECTRONICS GMBH & CO. KG. Invention is credited to Martin Liess, Juergen Schilz.
Application Number | 20090219971 12/065240 |
Document ID | / |
Family ID | 37247528 |
Filed Date | 2009-09-03 |
United States Patent
Application |
20090219971 |
Kind Code |
A1 |
Liess; Martin ; et
al. |
September 3, 2009 |
METHOD AND APPARATUS FOR CORRECTING THE OUTPUT SIGNAL OF A
RADIATION SENSOR AND FOR MEASURING RADIATION
Abstract
A method for correcting the output signal of a radiation sensor
20 includes obtaining two or more temperature signals from a
corresponding number of measurements of quantities at different
times and/or different locations relating to the temperature of the
sensor, and correcting the output signal with reference to said
temperature signals.
Inventors: |
Liess; Martin; (Wiesbaden,
DE) ; Schilz; Juergen; (Niedernhausen, DE) |
Correspondence
Address: |
PERMAN & GREEN
425 POST ROAD
FAIRFIELD
CT
06824
US
|
Assignee: |
PERKINELMER OPTOELECTRONICS GMBH
& CO. KG
Wiesbaden
DE
|
Family ID: |
37247528 |
Appl. No.: |
12/065240 |
Filed: |
August 28, 2006 |
PCT Filed: |
August 28, 2006 |
PCT NO: |
PCT/EP06/08402 |
371 Date: |
October 17, 2008 |
Current U.S.
Class: |
374/133 |
Current CPC
Class: |
G01J 5/16 20130101 |
Class at
Publication: |
374/133 |
International
Class: |
G01J 5/16 20060101
G01J005/16 |
Foreign Application Data
Date |
Code |
Application Number |
Aug 30, 2005 |
DE |
10 2005 041 050.2 |
Claims
1. A method for correcting the output signal of a radiation sensor,
comprising the steps of obtaining two or more temperature signals
from a corresponding number of measurements of quantities relating
to the temperature of the sensor or relating to one or more
components of the sensor, and correcting the output signal with
reference to said temperature signals.
2. The method according to claim 1, wherein correcting the output
signal is made also with reference to one or more calibration
values.
3. The method according to claim 1, in which the measurements are
spaced in time.
4. The method according to claim 3, wherein the time difference is
selected in accordance with the time constant of the sensor or of a
component thereof.
5. The method according to claim 1, in which the measurements are
spaced in locus.
6. The method according claim 1, comprising the steps of
determining a correction value with reference to said temperature
signals, and correcting the output signal with reference to said
correction value.
7. The method according to claim 1, in which a difference value of
two temperature signals or of at least one derived value derived
from said temperature signals is formed and used for
correction.
8. The method of claim 7, in which the derived value is an average
value.
9. The method of claim 8, wherein the average value is determined
as follows: va=k*Ta+(1-k)*vae, wherein va is the average value, vae
is an earlier corresponding average value, Ta is the actually
measured temperature value, and k is an averaging coefficient with
0<k.ltoreq.1.
10. The method according to claim 9, in which two average values
are determined, said two average values having different averaging
coefficients, wherein for correction the difference value of said
two average values is formed.
11. The method according to claim 1, in which the temperature
signals are transmitted away from the sensor for external
correaction of the output signal of the radiation sensor.
12. The method according to claim 1, in which one or more of the
temperature signals or a derived value derived from said
temperature signals are stored in the sensor and the output signal
of the radiation sensor is corrected within the sensor and is
output therefrom.
13. The method according to claim 1, comprising one or more of the
following features: the sensor comprises a thermopile sensor
element, preferably having cold and warm contacts, the warm
contacts preferably located on a membrane, the sensor comprises an
ASIC for obtaining the two or more temperature signals, and/or for
correcting the output signal with reference to said temperature
signals, the sensor element is adapted to convert IR radiation into
an electric signal.
14. A method for measuring a temperature, comprising the steps of
obtaining an output signal from a radiation sensor, and correcting
said output signal with a method according to one or more of the
preceding method claims.
15. An apparatus for measuring radiation, comprising a sensor
element (10) for receiving radiation and transforming it into an
electrical output signal, and means (11, 21, 22, 24) for obtaining
two or more temperature signals from a corresponding number of
measurements of quantities relating to the temperature of the
apparatus.
16. The apparatus according to claim 15, comprising correcting
means (21, 50, 80) for correcting the output signal with reference
to said temperature signals.
17. The apparatus according to claim 1, comprising means (29a e)
for outputting the output signal and the obtained temperature
signals and/or one or more derived values derived from the
temperature signals.
18. The apparatus according to claim 1, comprising one or more
temperature sensors (11, 22, 24) for rendering said temperature
signals.
19. The apparatus according to claim 1, comprising interrogating
means (21) for repeatedly interrogating a temperature sensor for
obtaining said temperature signals.
20. The apparatus according to claim 1, comprising subtracting
means (53) for forming a difference value of two temperature
signals or of at least one derived value derived from said
temperature signals, and means (55) for correcting the output
signal with reference to said difference value.
21. The apparatus according to claim 20, comprising at least one
averaging means (60) for forming an average value as said derived
value.
22. The apparatus according to claim 21, in which the averaging
means comprises a register (62) for holding an earlier average
value, and a calculator (61, 63, 64) determining the actual average
value from the actual temperature signal and the earlier average
value.
23. The apparatus according to claim 1, comprising an ASIC (21) and
a sensor element (10).
24. The apparatus according to claim 1, comprising transformation
means (44) for determining the temperature from said corrected
signal.
25. The apparatus according to claim 1, wherein the apparatus is a
sensor (20) with a sensor element (10), an ASIC (21), a housing
(36), a radiation permeable window (37) and terminals (29).
26. The method according to claim 1, wherein the temperature
signals are acquired at two or more different locations and with
two or more different time references, and correction is made with
reference to said acquired temperature signals.
27. The method of claim 26, wherein differences are formed in a
pair-wise manner amongst the acquired temperature signals, and the
differences are added with a weighting applied to them.
28. The method of claim 26, wherein the acquired temperature
signals are added with a weighting applied to them.
29. The method according to claim 1, wherein at least one
temperature signal relating to the temperature of the sensor or
relating to one or more components of the sensor is obtained from a
measurement outside the sensor.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is the National Stage of International
Application No. PCT/EP2006/008402, International Filing Date, Aug.
28, 2006, which designated the United States of America, and which
international application was published under PCT Article 21(2) as
WO Publication No. WO 2007/025697 A1 and which claims priority from
German Application No. 10 2005 041 050.2, filed Aug. 30, 2005.
BACKGROUND
[0002] 1. Field
[0003] The disclosed embodiments relate to a method and an
apparatus for correcting the output signal of a radiation sensor
and for measuring radiation. Related disclosures can be found in DE
102 004 028 032.0 and DE 102 004 028 022.3.
[0004] 2. Brief Description
[0005] Radiation sensors transform electromagnetic radiation into
an electrical signal. This may be accomplished, for example, by
thermopiles, bolometers or the like. The radiation sensed by them
is often infrared radiation (wavelength larger than 800 nm).
Radiation sensors of this type are often used for contactless
temperature measurement. The body of which the temperature is to be
measured emits radiation in dependence of its temperature. The
radiation is the more intense the higher the temperature of said
body is. Accordingly, the emitted infrared radiation of a body may
be used for contactless measuring its temperature. The details
thereof will be explained with reference to FIG. 1.
[0006] FIG. 1 shows a sensor element 10. It comprises a frame 2
which is a support for a membrane 3. The frame 2 surrounds an
opening 4 which may have rectangular or round cross-section
depending on particular necessities. The membrane 3 serves to
thermally insulate the actual sensing portion 1 formed on the top
surface of the membrane 3 from the surrounding as far as possible.
From the top surface, the sensing portion 1 of the sensor element
10 receives radiation, preferably infrared radiation, as indicated
by two arrows IRn and IRs. IRs indicates the desired signal
radiation from the body to be measured. However, the sensing
portion receives also noise radiation, as indicated by arrow IRn.
This may come from components in the immediate vicinity of the
sensing portion, for example the housing of the sensor, shielding
members, or the like. The sensing portion 1 itself cannot
distinguish which kind of radiation impinges on its surface. It
will transform both of them into an electrical signal.
[0007] If the sensing portion 1 comprises a thermopile consisting
of a sequence of hot and cold contacts, then the measurement
principle is that the incident radiation will transform into a
temperature change (usually rise of temperature) at the hot
ends/contacts 1a. In FIG. 1, the ends above the opening 4 are the
hot ends 1a of the thermopile, whereas the ends above the frame 2
are the cold ends lb. For enhancing measurement sensitivity, the
hot and cold ends may be covered with auxiliary layers,
particularly an absorbing layer 5 above the hot ends la and a
reflecting layer 6 above the cold ends lb. The incident radiation
causes a difference in temperature between the hot and the cold
ends, and in dependence of this temperature difference, the
thermopile will generate an electrical signal.
[0008] Another noise source is indicated by the thick arrow Ta. It
is heat conduction through the various physical bodies. 7 is a
substrate such as a silicon wafer, a ceramics baseboard or a
printed circuit board on which the sensor element 10 of FIG. 1 is
mounted. Changes in the ambient temperature will communicate
through heat conduction through the support 7, frame 2, and
membrane 3 to the sensing portion 1. Heat conduction also takes
place between the surrounding atmosphere and the sensor element 10
and the sensing portion 1 thereof, but heat conduction through the
substrate 7 is usually much stronger in effect. Since the cold ends
are usually differently located with respect to the frame 2 as the
warm ends, the former will experience a change in ambient
temperature earlier than the warm ends. The hot contact on the
membrane of the sensor element is usually the last relevant
component that experiences a temperature change because it is
usually the thermally best isolated part of the relevant
measurement system.
[0009] Thus, a change in ambient temperature will first be
experienced by the cold ends and only later by the warm ends of the
sensing portion 1. Accordingly, through heat conduction a
temperature difference builds up between the hot and the cold ends
which has nothing to do with the temperature difference caused by
the signal infrared radiation. The temperature difference caused by
heat conduction will be the larger the faster the temperature
change is, because in a fast transition through a temperature range
the sensor element will not go through the temperature range in a
state close to thermal equilibrium. It will not have almost the
same temperature everywhere on the sensor.
[0010] Rather, there will be temperature differences between the
hot and the cold ends which serve to cause errors in the output
signal and accordingly in the measured temperature.
[0011] The above two mentioned German patent applications of the
same applicant propose various ways for overcoming erroneous
measurements caused by temperature shocks of the ambience.
[0012] One proposal is to equalize the thermal flow towards the hot
and the cold ends by arranging them suitably with respect to the
frame 2 on the one hand side, and on the other hand side by
appropriately designing the auxiliary layers 5 and 6 (absorbing
layer, reflecting layer). However, in various applications this
cannot fully eliminate erroneous measurement. In many cases, it is
desired to have the cold ends above frame 2 because it serves as a
thermal mass and has the effect of keeping the cold ends at a
steady temperature when measurement is made. Accordingly, there is
a systematic desire for an asymmetric arrangement of the hot and
cold ends with respect to the frame 2, and the design of the
auxiliary layers cannot fully compensate this for changes of the
ambient temperature.
[0013] Another proposal is to design the housing of the sensor
element 10 such that noise radiation as symbolized by arrow IRn is
blocked from the sensing portion as far as possible.
[0014] But while the above proposals have significant advantageous
effects particularly by appropriately designing the components that
are needed anyway (sensor element 10 including frame, membrane,
thermopile, auxiliary layers, and also the housing of the sensor),
there are nevertheless situations where an even more sophisticated
compensation of error sources particularly at changing ambient
temperature ("thermal shock") is desired.
SUMMARY
[0015] It is the object of the disclosed embodiments to provide a
method and an apparatus for correcting the output signal of a
radiation sensor and for radiation measurement with high
accuracy.
[0016] This object is accomplished in accordance with the feature
of the independent claims. Dependent claims are directed on
preferred embodiments of the disclosed embodiments.
[0017] A method of correcting the output signal of a radiation
sensor comprises the steps of obtaining two or more temperature
signals from a corresponding number of measurements of quantities
relating to the temperature of the radiation sensor, and correcting
the output signal with reference to said temperature signals.
[0018] A method for measuring the temperature of an object
comprises the steps of obtaining an output signal from a radiation
sensor receiving radiation from said object in accordance with said
radiation impinging on said sensor, and correcting the output
signal with a method as mentioned above.
[0019] An apparatus for measuring radiation comprises a sensor
element for receiving radiation and transforming it into an
electrical output signal, and means for obtaining two or more
temperature signals from a corresponding number of measurements of
quantities relating to the temperature of the apparatus. Said two
or more temperature signals are used for correcting the output
signal. The temperature can be determined from said corrected
output signal.
[0020] According to the disclosed embodiments, two or more
temperature measurements of the temperature of the sensor or the
sensor element or the sensing portion are obtained for obtaining a
measure for the thermal imbalance. The two or more temperature
measurements may be spaced in locus and/or spaced in time. In any
case, they will reflect thermal dynamics relating to the sensor
temperature and allow conclusions relating to the thermal imbalance
caused to the hot contacts 1a and the cold contacts 1b of the
sensing portion 1.
[0021] In an appropriate evaluation mechanism, these temperature
measurements can be evaluated by providing correction values for
the output signal from said temperature measurements, and/or by
immediately correcting the output signal of the sensor element with
reference to said temperature measurements.
BRIEF DESCRIPTION OF THE DRAWINGS
[0022] In the following, embodiments of the disclosed embodiments
will be described with reference to the attached drawings, in
which
[0023] FIG. 1 is a schematic sectional view of an embodiment of the
sensor element incorporating features of the disclosed
embodiments,
[0024] FIG. 2 is a schematic plain view of a sensor formed in
accordance with an embodiment of the disclosed embodiments,
[0025] FIG. 3 is a schematic sectional view of an apparatus for
measuring the temperature,
[0026] FIG. 4 is a schematic structure of the signal
processing,
[0027] FIG. 5 is a block diagram of a way of processing a
temperature signal in a correcting means, and
[0028] FIG. 6 is a block diagram of how to obtain a particular
average value,
[0029] FIG. 7 is a diagram showing typical signal curves,
[0030] FIG. 8 is a schematic representation of another correcting
means,
[0031] FIG. 9 is a schematic representation of yet another
correcting means.
DETAILED DESCRIPTION
[0032] FIG. 1 shows a sectional view of a sensor element 10 formed
in accordance with an embodiment of the disclosed embodiments. 2 is
a frame formed by micromachining, for example from a silicon wafer.
It may have a rectangular outer cross section. An opening 4 with
rectangular or partially or fully rounded cross section is
surrounded by the frame 2. A membrane 3 spans across the opening 4.
On the membrane 3, the sensing portion 1 is formed. It may be a
thermopile with a couple of warm contacts 1a and cold contacts 1b.
The warm contacts 1a are usually located above the opening 4. The
cold contacts may be located above the frame 2 or also above the
opening 4, depending on particular necessities. In measurement, the
warm contacts 1a have a temperature T2, whereas the cold contacts
1b have a temperature T1. From said temperature difference, the
actual electrical signal will be determined. An absorbing layer 5
for enhancing absorption may be provided above the warm contacts
1a, and a reflecting layer 6 for preventing absorption may be
provided above the cold contacts 1b.
[0033] According to one embodiment of the disclosed embodiments,
one or more temperature sensors 11 may be provided on the sensor
element 10. They may be provided on an arbitrary position of the
sensor element 10, but preferably distant from the hot contacts 1a,
e.g. close to the cold contact 1b and/or inbetween cold contact 1b
and warm contact 1a.
[0034] For describing one embodiment of signal evaluation of the
disclosed embodiments, it is in the following assumed that one
temperature sensor 11 is provided close to the cold contacts 1b and
another one is provided in between cold and warm contacts, as shown
in FIG. 1. If, caused by a thermal shock, a temperature change
sweeps through the sensor element 10 as indicated by thick arrow
Tn, this will first by experienced by the cold contact 1b and by
the accordingly allocated temperature sensor 11a, and thereafter it
will be experienced by the temperature sensor 11b located between
warm and cold contacts. Accordingly, the two temperature sensors
will show different temperatures, they show a gradient over locus.
This gradient is not caused by the radiation to be measured.
Rather, it reflects the thermal shock experienced by the sensor
element 10 and particularly, the thermal imbalance (noise
imbalance) caused by the change of ambient temperature in addition
to the thermal imbalance (signal imbalance) caused by the infrared
radiation from the objects to be measured.
[0035] A temperature sensor 11 may have own electrical terminals
through which its signal can be interrogated. It can be, for
example, a temperature resistant resistor or similar devices.
[0036] The above embodiment measures the temperature at two
locations on the sensor element 10, itself. However, it is not
necessary to measure the thermal imbalance immediately at the
sensor element itself. Rather, it may also be measured between the
sensor element 10 and another component, for example the substrate
7 because also such an imbalance is a measure for the thermal
inequilibrium caused by change in ambient temperature (thermal
shock). Accordingly, there need not be two sensor elements provided
on the sensor element 10 itself. Rather, one may be provided
somewhere on the sensor element 10, and another in another
component of the sensor.
[0037] FIG. 2 shows an embodiment of a sensor 20 in a schematically
way. It is a plain view on a base plate of a possibly housed sensor
20 with the housing, for example a cap member, being removed. 10
symbolizes the sensor element of FIG. 1 with a temperature sensor
11 thereon. 21 symbolizes an evaluation electronics which may be an
ASIC (application specific integrated circuit). 29a to e symbolize
contact points for sensor terminals. Not shown is a wiring between
sensor element 10, evaluation electronics 21 and contacts 29a to e.
22 symbolizes a temperature sensor on the base plate of the sensor
20, said base plate having reference numeral 25 in FIG. 2. It may
be component 7 in FIG. 1. The evaluation electronics 21 may itself
have a temperature sensor 24 formed thereon.
[0038] In the embodiment shown in FIG. 2, at least two of the
temperature sensors 11, 22 and 24 may be used. They are provided on
suitable differing locations on the sensor 20, and they will show a
temperature gradient over locus not being caused by the signal
infrared radiation to be measured, but by a change of ambient
temperature. Again, such a gradient can be used for correcting the
output signal of the sensor element 10. For evaluation, one may for
example consider the temperature difference between sensor elements
24 and 11, or between 22 and 11, or between 24 and 22. In the later
option, it is not at all necessary to provide a temperature sensor
on the sensor element 10 itself.
[0039] In an embodiment of the disclosed embodiments, the apparatus
for measuring radiation may comprise only a sensor as schematically
shown in FIG. 2, said sensor having the sensor element 10 and means
21 for correcting the output signal. Said means 21 for correcting
the output signal may be an ASIC formed within sensor 20. ASIC 21
receives the raw output signal of sensor element 10, obtains the
temperature measurements, and corrects the raw output signal of
sensor element 10 and outputs the correct signal to the terminals
29a to e.
[0040] At least one temperature signal relating to the temperature
of the sensor or to one or more components of the sensor and used
for correction may be obtained from a measurement outside the
sensor, for example a measurement on the circuit board where the
sensor is mounted. The signal may then be inputted to the sensor in
an appropriate manner or it may be used outside the sensor on or
with quantities output from the sensor.
[0041] In another embodiment of the disclosed embodiments, the
apparatus for measuring radiation may be a larger system in which
the raw signal from the sensor element 10 (perhaps amplified and
calibrated in sensor 20) is transmitted away from the sensor 20
towards an external circuit for further processing there.
[0042] The sensor element 10 may have a size of less than 3 mm*3
mm, preferably less than 2 mm*2 mm. The sensor 20 may have a
regular or standardized housing such as a TO5-housing. Multiple
sensor elements 10 may be provided in one sensor. Each output
signal thereof may be corrected as described. Signal multiplexing
may be used for this as well as for signal output.
[0043] FIG. 3 shows an embodiment of an electronic component which
schematically shows in cross section a housed circuit. 31 is a
baseboard, for example a printed circuit board. 32 may be a socket
for a radiation sensor. 20 symbolizes the radiation sensor itself
in the side view, it shows the sensor base plate 25, a cap 36
housing and closing the sensor, a radiation entrance window 37
which may comprise a focusing element such as a lens or a mirror,
and terminals 38 received by the socket 32 or immediately soldered
to the circuit board 31. 39a to c symbolize other circuit elements
such as resistors, capacitors, and the like. 33 may be again an
ASIC or a digital component such as a microprocessor. 35 symbolizes
a connector for transmitting away signals and receiving signals and
for power supply.
[0044] The temperature signals obtained from the at least two
measurements within sensor 20 may be transmitted away from sensor
20 together with the raw (and possibly amplified and calibrated)
output signal of the sensor element 10. These signals may be
processed for example in ASIC or microprocessor 33, and corrected
values are further used or outputted via connector 35.
[0045] In yet another, not shown embodiment, circuit 30 as shown in
FIG. 3 may also be some kind of preprocessing, signal formatting
and process control, and signals corresponding to the temperature
measurements and the raw output signal (perhaps calibrated and
amplified) of the sensor element 10 are transmitted away from
circuit 30 towards a regular computer for further processing
there.
[0046] In the following explanations, it is assumed that the entire
correction is made within sensor 20 of FIG. 2. However, as
indicated above, it may also be made in external components.
[0047] FIG. 4 shows a general signal flow. 10 indicates the sensor
element, which outputs a raw signal (voltage) Vr. This signal Vr
may be amplified in an amplifier 42 giving an amplified voltage Va,
which may further linearly be calibrated for offset and sensitivity
in a calibration 43, this giving a calibrating voltage Vc. A
transformation means 44 transforms the calibrating voltage Vc into
a voltage reflecting the temperature Vt of the object to be
measured. Preferably, prior to transformation means 44, the
correaction of the obtained signals as described above is made. In
FIG. 4, this is schematically shown by box 21 representing the
correcting means 21 as shown in FIG. 2, which may be the ASIC
within sensor 20 or an external component as shown with reference
numeral 33 in FIG. 3 or a (not shown) usual computer.
[0048] The correction means 21 may be preferably inserted between
sensor element 10 and amplifier 42 or between amplifier 42 and
calibration 43 or between calibration 43 and transformation means
44. The transformation means may involve Botzmanns T 4 dependency.
Correcting means 21 receives the uncorrected (but perhaps already
amplified and/or calibrated) signal, corrects it as mentioned above
in accordance with the at least two measurements of temperature of
the sensor or a particular component thereof, and outputs it for
further processing. Correction means 21 receives the at least two
temperature measurements Tn and Tm as indicated with boxes 45 and
46 and may further receive calibration values 47.
[0049] The correction may be performed on the analog or on the
digital side. Likewise, calibration 43 may be analog or
digital.
[0050] Amplification 42 and calibration 43 may be performed in a
unified component or may be reversed in order as compared to what
is shown in FIG. 4. Likewise, one or more of the boxes 42, 43 and
44 may be incorporated in the correcting means 21 to form a unified
piece of hardware such as the mentioned ASIC.
[0051] So far, temperature gradients over locus were described. In
another embodiment of the disclosed embodiments, a temperature
gradient over time is obtained. It may then not be necessary to
obtain temperature measurements at two or more locations. This
embodiment reflects the fact that a temperature gradient in time
correlates strongly with a temperature gradient over locus. Looking
at the entire measuring apparatus when it experiences a temperature
shock, this shock will cause a temperature gradient over locus with
the peripheral components experiencing the temperature change
first, and more central components experiencing the temperature
change later, thus rendering a gradient over locus, as explained
above. By the way, the innermost component in this respect will
usually be the hot contact on the membrane of the sensor element,
because usually this is the thermally best isolated part of the
relevant measurement system.
[0052] However, looking at one particular locus of the sensor 20 as
shown in FIG. 20 or sensor element 10 as shown in FIG. 1, this
locus will almost always also experience a temperature gradient
over time when a temperature shock through change of ambient
temperature is experienced. As long as the entire measurement
system is in thermal equilibrium, its components have the same
temperature and won't show a gradient over locus, and their
temperature is stable and won't show a gradient over time, either.
However, if a temperature shock is experienced, this will lead both
to temperature changes at a particular location and thus giving a
temperature gradient over time there, until the new thermal
equilibrium is reached, so that also a thermal gradient from two or
more temperature measurements spaced in time is suitable for
detecting the circumstances that may lead to a temperature
difference at hot and cold contacts 1a and 1b of the sensing
portion 1 of the sensor element 10 in FIG. 1. Then, only one
temperature sensor of those in FIG. 1 may be sufficient, for
example sensor 11 provided on the sensor element 10, or sensor 22
provided on the base plate of the radiation sensor, or sensor 24
provided in the correcting means, for example the ASIC. It is
justified to assume that in practically all applications the
temperature changes of the individual components of the overall
sensor 20 will not grossly deviate from each other. Rather, they
will be similar. Therefore, measuring a temperature gradient over
time at a locus different from the sensor element 10 itself quality
reflects the circumstances requiring the correction according to
the disclosed embodiments.
[0053] In the above, one embodiment was described in which a
gradient over locus was obtained, and another embodiment was
described, in which a gradient over time was obtained. Generally
speaking, locus dependent measurement and time dependent
measurement can be combined to evaluate temperature differences
both over time and over locus. All these values may then be used
for appropriate correction in the correction means 21.
[0054] Generally speaking, one way of providing correction to the
uncorrected signal is to form a difference between at least two of
the obtained temperature values and to apply a correction
proportional to the difference additively or multiplicatively to
the uncorrected signal. Instead of the temperature values used for
forming the above-mentioned difference, values derived from said
temperature values may be used, particularly average values.
Averaging has the advantage that the useful signal will sum up,
whereas noise tends to neutralize itself. Averaging may be
particularly used if the gradient over time of the temperature
signal is evaluated. Particularly, an auto-regressive average of
temperature values measured over time may be acquired according to
the formula
va=k*Ta+(1-k)*vae,
[0055] wherein va is the average value to be determined, vae is an
earlier corresponding average value, Ta is the actually measured
temperature value, and k is an averaging coefficient between 0 and
1. The value k is a weighting coefficient that weights the present
temperature value Ta in relation to the value vae incorporating the
earlier values of Ta. Together, the entire weight is 1. If k is
large, then the actual temperature strongly influences the new
average value va and the earlier composite value vae has weaker
influence thereon, whereas when k is small, the actual temperature
Ta only weakly influences va whereas the earlier values
incorporated in vae have stronger effect thereon. Therefore, by
setting k, one can determine whether the effective time of the
average value va is closer to the present or closer to the past. In
the extreme, if k is 1, then the history incorporated in the
earlier value has no influence at all, because it is multiplied
with zero.
[0056] If temperature values of different times are desired for
obtaining the gradient over time, then one may use two auto-average
values as indicated above with different averaging coefficients k
such that the one of them is closer to the present value and the
other is stronger adhered to the past.
[0057] The value k can be selected in view of the time constant of
the sensor element 10 (more in detail: the time constant for the
hot contacts to react on the temperature change applied through
heat conduction from the bottom of the frame). Further, the
averaging parameter k can be selected in accordance with the
sampling rate of the device 21 performing the correction. And
further, the sampling rate can be determined in accordance with
said time constant.
[0058] FIG. 5 shows a block diagram of a correcting means 50 for
performing the immediate correction. The correcting means may be
part of correcting means 21. Its input signals Ta, Ts and output
signal Tk may be one or more of the values Tn, Tm, Vr, Va, Vc or Vt
in FIG. 4. It receives the signal from the sensor element 10,
symbolized as signal Ts in FIG. 5, which may have undergone already
some preferably linear processing such as amplification and/or
calibration as shown in FIG. 4. Further, correcting means 50
receives signal Ta representing the measured temperature measured
by one temperature sensor such as one of reference numerals 11, 22
and 24 in FIG. 2. Register 51 keeps an actual value, and register
52 keeps a past value. 53 is a subtractor in which the earlier
value from 52 is subtracted from the later value at 51. The
difference goes to a calibration 54 which may perform a preferably
linear correction. Then it is applied to the uncorrected
temperature signal Ts in box 55. It may be an addition or a
multiplication or some kind of nonlinear correction in accordance
with the calibrated value leaving box 54. For example, a table may
be addressed, the table outputting correction values for correcting
Ts. The thus corrected signal Ts leaves box 55 and the correcting
means 50 as a signal Tk for further processing, particularly for
sooner or later entering box 44 in FIG. 4.
[0059] In an embodiment, the value from register 51 is transferred
to register 52 after the difference of the registered values was
formed, register 51 receives a new value of Ta, and the procedure
starts again.
[0060] So far, with reference to FIG. 5 a procedure was described
in which immediate temperature values Ta were used for correction.
However, as said above, one or more derived values (derived from
the temperature signal Ta) may be used instead. FIG. 6 shows a
block diagram of an averaging means 60 that may be used for example
as block 51 in FIG. 5 and/or as block 52 in FIG. 5. It forms an
autoregressive average as mentioned above. 62 is a register holding
a value. Ta is the input of the measured temperature. 61 symbolizes
a multiplier to multiply the input value with the averaging
coefficient k (0<k.ltoreq.1), and the result goes to an adder 64
which also receives the content of register 62 multiplied by 1-k in
multiplier 63. The sum of both is again written to register 62 and
output as an average value va.
[0061] Using an autoregressive average has the advantage that not a
plurality of registers is necessary for holding past values.
Rather, said past values are all contained in the already held
average value which is added to the appropriately weighted new
temperature value for registering in the same register as the
earlier value by overwriting it.
[0062] In FIG. 5, both registers 51 and 52 may be replaced by
respectively one averager 60 as shown in FIG. 6, but these
averagers working with differing averaging coefficients. The one in
the top of FIG. 5 has a higher coefficient (closer to 1) and is
thus closer to the actual value of Ta, whereas the lower thereof
has a smaller value of k (closer to 0) so that its output is closer
to the past. Instead of receiving the same inputs Ta from one
temperature sensor, such averagers 60 may receive differing inputs
from differing temperature sensors as shown in FIG. 2. If they
receive different temperature inputs, they may have the same
averaging coefficient k.
[0063] The result of the FIG. 5 correcting means 50 using two
averagers 60 as shown in FIG. 6 with differing averaging
coefficients k on the same input Ta is shown in FIG. 7. The curve
T(t) symbolizes a temperature change in the temperature as
experienced by a temperature sensor 1, 22, 24 as shown in FIG. 2.
Curve va1(t) symbolizes the autoregressive average with a higher k
(i.e. quicker following T(t)), whereas va2(t) represents the curve
of the autoregressive average having a smaller k (thus following
curve T(t) slower). If one looks at the respective average values
at a particular point of time tx, then it shows that curve va1(t)
has a value of the curve T(t) at point of time tf, whereas the
slower curve va2(t) has a value of an earlier point of time ts.
Accordingly, with the two averaging coefficients k used in the two
averagers 60 one can determine to which extent the two obtained
average values deviate as regards their effective time.
[0064] As long as two values of measured temperature or values
derived therefrom are taken, only one difference thereof can be
formed. This difference can be set appropriately by numerically
adjusting the averaging coefficients used in boxes 61 and 63,
respectively, and also by adjusting the coefficients in boxes 42,
43 and 54. However, it is also possible to use more than two
temperature values or more than two derived values derived from
temperature values. In a preferred embodiment, the temperature of
the sensor 20 or sensor element 10 may be measured at two or more
different locations and with two or more different time references,
such as different points of time of different measurements or
different effective times of different autoregressive averages as
mentioned above. One obtains then at least four values which allow
formation of at least six differences amongst them. FIG. 8 shows a
corresponding embodiment.
[0065] 80 is a correcting means functionally corresponding to
correcting means 50 in FIG. 5. It receives a signal Ta1
representing the temperature at a first location, and a signal Ta2
representing the temperature at a second location. Both signals
respectively may go through a fast and a slow autoregressive
averaging process as described with reference to FIG. 6, thus
rendering four values relating to different locations at different
times. Instead of the averagers 60, also storage registers may be
used with an appropriate renewal structure behind them.
[0066] Accordingly, four values are available for forming
differences amongst them at subtractors 81, these differences
reflecting a gradient over locus and/or a gradient over time. In a
calibration process, there may coefficients 82 for each of said
differences be determined for properly taking into account said
difference for correcting the temperature signal to be corrected Ts
from the sensor in order to produce the corrected temperature
signal Tk. This may be accomplished in a calibration process in
which a sensor in its built-in state is exposed to a defined change
of ambient temperature so that the respective sensor signals are
obtained (Ts from the radiation sensor on the one hand side and
Ta1, Ta2 at least on the other hand side). By a heuristic
optimization process performed by numerically processing and
comparing the respective data, coefficients 82 for the respective
differences can be obtained and permanently stored in the
correcting means 80 as shown in FIG. 8, preferably by writing them
into PROM-like registers. The weighted differences may be added in
an adder 83 and used for correction of Ts in box 84 to obtain
Tk.
[0067] Generally speaking, coefficients used in the above described
techniques may be obtained by calibrating an individual sensor,
possibly in its built-in state, in a defined environment in which
the respective outputs are monitored and the coefficients are set
such that deviation between actual and target values become
minimum. Coefficients may be permanently written into the sensor,
e.g. into the correcting means 21.
[0068] Instead of the structure of FIG. 8 also the one in FIG. 9
can be used. Behind this is the idea that the differences formed in
FIG. 8 are, and go through, linear operations so that instead of
separately forming and weighting the differences and adding them,
also their input values can be weighted and added. Each value
leaving the boxes 60 in FIG. 8 contributes to three differences
either on the (+)-side thereof or on the (-)-side. Assume that one
of the values is in two differences weighted with 0.20 and 0.14,
respectively, on the (+)-side, and in one difference weighted with
0, 15 on the (-)-side. Then its entire weight in the final result
is 0.20+0.14-0.15=0.19. A thus obtained weighting coefficient may
be negative. This weight can be applied to the output of averagers
60 or to corresponding values, as shown in FIG. 9, and the weighted
results are summed up. From a computational point of view, this is
less complex than the embodiment in FIG. 8 and renders the same
result.
[0069] For using the at least two temperature measurements for
correction purposes, one can evaluate them in any suitable manner
for obtaining a correction reflecting the temperature dynamics
experienced by the sensing portion 1 of the sensor element. So far,
subtractions were described as evaluation (reference numerals 53,
83). But other evaluations may be used instead for rendering
results reflecting said temperature dynamics and particularly a
noise temperature difference as described with reference to FIG.
1.
[0070] For properly performing the respective tasks, the correcting
means 21 in FIG. 2 may have one or more clocked tasks which are
repeatedly executed. All required processings may be compiled to
one big task executed with a suitable repetition rate. Such a task
may comprise data acquisition (from at least of sensor element 10,
and from one or more temperature sensors 11, 22, 24), calibration,
subtraction, and the like, as described above.
* * * * *